--- a/progs/re4.scala Wed Oct 02 00:10:52 2013 +0100
+++ b/progs/re4.scala Wed Oct 02 01:02:50 2013 +0100
@@ -1,4 +1,5 @@
import scala.annotation.tailrec
+
abstract class Rexp {
def simp : Rexp = this
}
@@ -12,7 +13,7 @@
case (r, NULL) => r
case (r, EMPTY) => if (nullable(r)) r else ALT(r, EMPTY)
case (EMPTY, r) => if (nullable(r)) r else ALT(r, EMPTY)
- case (r1, r2) => ALT(r1, r2)
+ case (r1, r2) => if (r1 == r2) r1 else ALT(r1, r2)
}
}
case class SEQ(r1: Rexp, r2: Rexp) extends Rexp {
@@ -24,8 +25,21 @@
case (r1, r2) => SEQ(r1, r2)
}
}
-case class STAR(r: Rexp) extends Rexp
-case class NTIMES(r: Rexp, n: Int) extends Rexp
+case class STAR(r: Rexp) extends Rexp {
+ override def simp = r.simp match {
+ case NULL => EMPTY
+ case EMPTY => EMPTY
+ case r => STAR(r)
+ }
+}
+case class NTIMES(r: Rexp, n: Int) extends Rexp {
+ override def simp = if (n == 0) EMPTY else
+ r.simp match {
+ case NULL => NULL
+ case EMPTY => EMPTY
+ case r => NTIMES(r, n)
+ }
+}
// some convenience for typing in regular expressions
def charlist2rexp(s : List[Char]) : Rexp = s match {
@@ -45,7 +59,7 @@
case ALT(r1, r2) => nullable(r1) || nullable(r2)
case SEQ(r1, r2) => nullable(r1) && nullable(r2)
case STAR(_) => true
- case NTIMES(r, i) => if (i == 0) false else nullable(r)
+ case NTIMES(r, i) => if (i == 0) true else nullable(r)
}
// derivative of a regular expression w.r.t. a character
@@ -73,18 +87,10 @@
def matcher(r: Rexp, s: String) : Boolean = nullable(ders(s.toList, r))
-
//one or zero
def OPT(r: Rexp) = ALT(r, EMPTY)
-//n-times
-/*def NTIMES(r: Rexp, n: Int) : Rexp = n match {
- case 0 => NULL
- case 1 => r
- case n => SEQ(r, NTIMES(r, n - 1))
-}*/
-
-def RTEST(n: Int) = SEQ(NTIMES(OPT("a"), n), NTIMES("a", n))
+def EVIL(n: Int) = SEQ(NTIMES(OPT("a"), n), NTIMES("a", n))
def time_needed[T](i: Int, code: => T) = {
val start = System.nanoTime()
@@ -94,8 +100,8 @@
}
-for (i <- 1 to 13001 by 500) {
- println(i + " " + time_needed(1, matcher(RTEST(i), "a" * i)))
+for (i <- 1 to 12001 by 500) {
+ println(i + " " + "%.5f".format(time_needed(1, matcher(EVIL(i), "a" * i))))
}